DOP 238: Unlocking the Potential of Modern Architectures Using Service Mesh

#238: You might just now be reaching the point of containerizing your workloads and running them in Kubernetes. However, how do you go about accessing those workloads? You might have started with an ingress, because other options seemed like they were too complex.
However, you've now crossed the point where you need that complexity. That's where service mesh can help you out.
On this episode, we speak with Marino Wijay, Principal Developer Advocate at solo.io, about how service mesh has evolved over time, the challenges of adopting certain models, and the advantages of open-source solutions. We also talk about the differences between sidecar and sidecar-less models, and the circumstances in which either might be appropriate.
Marino's contact information:
Twitter: / virtualized6ix
LinkedIn: / mwijay
CZcams channel:
/ devopsparadox
Books and Courses:
Catalog, Patterns, And Blueprints
www.devopstoolkitseries.com/p...
Review the podcast on Apple Podcasts:
www.devopsparadox.com/review-...
Slack:
www.devopsparadox.com/slack/
Connect with us at:
www.devopsparadox.com/contact/

[@fenarRH]

This is a critical topic for holistic observability and traffic steering. We have tried using (telco network intensive workloads) a sidecar approach that brought overhead traffic bottlenecking, lcm issues of another sw piece, another attack surface, etc. Now, ambient mesh seems to address most of these with the cost of bringing overlay networking, which we do not necessarily like/love on cloud infrastructures. We start seeing use of ebpf for service information collection, traffic steering in this domain now. I would love to see a really round table discussion on this matter as it is very important for many service providers.